kernel/linux.git/include/linux/u64_stats_sync.h, branch v6.6.132

netfilter: nft_counter: fix reset of counters on 32bit archs

2026-03-04T12:19:51+00:00

[ Upstream commit 1e13f27e0675552161ab1778be9a23a636dde8a7 ] nft_counter_reset() calls u64_stats_add() with a negative value to reset the counter. This will work on 64bit archs, hence the negative value added will wrap as a 64bit value which then can wrap the stat counter as well. On 32bit archs, the added negative value will wrap as a 32bit value and _not_ wrapping the stat counter properly. In most cases, this would just lead to a very large 32bit value being added to the stat counter. Fix by introducing u64_stats_sub(). Fixes: 4a1d3acd6ea8 ("netfilter: nft_counter: Use u64_stats_t for statistic.") Signed-off-by: Anders Grahn Signed-off-by: Florian Westphal Signed-off-by: Sasha Levin

u64_stats: fix u64_stats_init() for lockdep when used repeatedly in one file

2024-04-17T09:19:28+00:00

[ Upstream commit 38a15d0a50e0a43778561a5861403851f0b0194c ] Fix bogus lockdep warnings if multiple u64_stats_sync variables are initialized in the same file. With CONFIG_LOCKDEP, seqcount_init() is a macro which declares: static struct lock_class_key __key; Since u64_stats_init() is a function (albeit an inline one), all calls within the same file end up using the same instance, effectively treating them all as a single lock-class. Fixes: 9464ca650008 ("net: make u64_stats_init() a function") Closes: https://lore.kernel.org/netdev/ea1567d9-ce66-45e6-8168-ac40a47d1821@roeck-us.net/ Signed-off-by: Petr Tesarik Reviewed-by: Simon Horman Reviewed-by: Eric Dumazet Link: https://lore.kernel.org/r/20240404075740.30682-1-petr@tesarici.cz Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

u64_stat: Remove the obsolete fetch_irq() variants.

2023-01-13T04:57:43+00:00

There are no more users of the obsolete interface u64_stats_fetch_begin_irq() and u64_stats_fetch_retry_irq(). Remove the obsolete API. [bigeasy: Split out the bits from a larger patch]. Signed-off-by: Thomas Gleixner Signed-off-by: Sebastian Andrzej Siewior Link: https://lore.kernel.org/r/20230110160738.974085-1-bigeasy@linutronix.de Signed-off-by: Jakub Kicinski

u64_stats: Streamline the implementation

2022-09-19T12:35:08+00:00

The u64 stats code handles 3 different cases: - 32bit UP - 32bit SMP - 64bit with an unreadable #ifdef maze, which was recently expanded with PREEMPT_RT conditionals. Reduce it to two cases (32bit and 64bit) and drop the optimization for 32bit UP as suggested by Linus. Use the new preempt_disable/enable_nested() helpers to get rid of the CONFIG_PREEMPT_RT conditionals. Signed-off-by: Thomas Gleixner Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra (Intel) Link: https://lore.kernel.org/r/20220825164131.402717-9-bigeasy@linutronix.de

u64_stats: Disable preemption on 32bit UP+SMP PREEMPT_RT during updates.

2021-12-13T12:42:08+00:00

On PREEMPT_RT the seqcount_t for synchronisation is required on 32bit architectures even on UP because the softirq (and the threaded IRQ handler) can be preempted. With the seqcount_t for synchronisation, a reader with higher priority can preempt the writer and then spin endlessly in read_seqcount_begin() while the writer can't make progress. To avoid such a lock up on PREEMPT_RT the writer must disable preemption during the update. There is no need to disable interrupts because no writer is using this API in hard-IRQ context on PREEMPT_RT. Disable preemption on 32bit-RT within the u64_stats write section. Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: David S. Miller

u64_stats: Introduce u64_stats_set()

2021-10-18T11:54:41+00:00

Allow to directly set a u64_stats_t value which is used to provide an init function which sets it directly to zero intead of memset() the value. Add u64_stats_set() to the u64_stats API. [bigeasy: commit message. ] Signed-off-by: Ahmed S. Darwish Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: David S. Miller

u64_stats,lockdep: Fix u64_stats_init() vs lockdep

2021-03-10T08:51:45+00:00

Jakub reported that: static struct net_device *rtl8139_init_board(struct pci_dev *pdev) { ... u64_stats_init(&tp->rx_stats.syncp); u64_stats_init(&tp->tx_stats.syncp); ... } results in lockdep getting confused between the RX and TX stats lock. This is because u64_stats_init() is an inline calling seqcount_init(), which is a macro using a static variable to generate a lockdep class. By wrapping that in an inline, we negate the effect of the macro and fold the static key variable, hence the confusion. Fix by also making u64_stats_init() a macro for the case where it matters, leaving the other case an inline for argument validation etc. Reported-by: Jakub Kicinski Debugged-by: "Ahmed S. Darwish" Signed-off-by: Peter Zijlstra (Intel) Tested-by: "Erhard F." Link: https://lkml.kernel.org/r/YEXicy6+9MksdLZh@hirez.programming.kicks-ass.net

u64_stats: Document writer non-preemptibility requirement

2020-06-04T22:50:42+00:00

The u64_stats mechanism uses sequence counters to protect against 64-bit values tearing on 32-bit architectures. Updating such statistics is a sequence counter write side critical section. Preemption must be disabled before entering this seqcount write critical section. Failing to do so, the seqcount read side can preempt the write side section and spin for the entire scheduler tick. If that reader belongs to a real-time scheduling class, it can spin forever and the kernel will livelock. Document this statistics update side non-preemptibility requirement. Reword the introductory paragraph to highlight u64_stats raison d'être: 64-bit values tearing protection on 32-bit architectures. Divide documentation on a basis of internal design vs. usage constraints. Reword the u64_stats header file top comment to always mention "Reader" or "Writer" at the start of each bullet point, making it easier to follow which side each point is actually for. Clarify the statement "whole thing is a NOOP on 64bit arches or UP kernels". For 32-bit UP kernels, preemption is always disabled for the statistics read side section. Signed-off-by: Ahmed S. Darwish Reviewed-by: Sebastian Andrzej Siewior Signed-off-by: David S. Miller

u64_stats: provide u64_stats_t type

2019-11-08T04:03:08+00:00

On 64bit arches, struct u64_stats_sync is empty and provides no help against load/store tearing. Using READ_ONCE()/WRITE_ONCE() would be needed. But the update side would be slightly more expensive. local64_t was defined so that we could use regular adds in a manner which is atomic wrt IRQs. However the u64_stats infra means we do not have to use local64_t on 32bit arches since the syncp provides the needed protection. Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller

net: u64_stats_sync: Remove functions without user

2018-05-08T03:25:24+00:00

Commit 67db3e4bfbc9 ("tcp: no longer hold ehash lock while calling tcp_get_info()") removes the only users of u64_stats_update_end/begin_raw() without removing the function in header file. Remove no longer used functions. Cc: Eric Dumazet Signed-off-by: Anna-Maria Gleixner Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: David S. Miller